DE69132889T2 - SUPERIOR THROMBOMODULIN ANALOGS FOR PHARMACEUTICAL USE - Google Patents

Abstract

The present invention relates to the use of analogs of thrombomodulin ("TM") that have the ability to enhance the thrombin-mediated activation of protein C but which have a significantly reduced ability to inhibit the direct procoagulant activities of thrombin, such as, for example, thrombin-mediated conversion of fibrinogen to fibrin. These analogs are useful in, for example, antithrombotic therapy. Novel proteins, nucleic acid gene sequences, pharmaceuticals and methods of inhibiting thrombotic activity are disclosed. Included are methods for increasing the circulating half life of the proteins.

Description

This application is a continuation-in-part of U.S. Serial No. 506,325, filed April 16, 1990, and PCT Serial No. 90/00955, filed February 16, 1990, which was a continuation-in-part of U.S. Serial No. 406,941, filed September 13, 1989, which was a continuation-in-part of U.S. Serial No. 345,372, filed April 28, 1989.

BACKGROUND OF THE INVENTION Field of the invention

The present invention relates to the use of thrombomodulin ("TM") analogues which have the ability to promote thrombin-mediated activation of protein C, but which have a significantly reduced ability to inhibit the direct coagulation-promoting effects of thrombin, such as the thrombin-mediated conversion of fibrinogen to fibrin. These analogues are useful, for example, in thrombosis therapy. New proteins, nucleic acid gene sequences, medicaments and methods for inhibiting thrombus formation are disclosed. Also disclosed are methods for extending the circulating half-life of the proteins.

There are many conditions that would benefit from treatment with a safe, effective anticoagulant/antithrombotic. These conditions are of various types. For example, therapy with anticoagulants is suitable for acute conditions such as thrombolytic therapy for myocardial infarction or for the treatment of disseminated intravascular coagulation (DIC) associated with, for example, blood poisoning. Anticoagulants are also suitable for less acute conditions, for example for chronic use in patients undergoing heart valve implantation or for prophylactic Use in outpatients to reduce the risk of deep venous thrombosis (DVT).

Information disclosure

Thrombomodulin is a membrane protein that has been shown to have anticoagulant properties. Its physiological significance has been investigated (see, for example, N. Esmon, et al., (1982) J. Biol. Chem. 257: 859-864, H. Salem, et al., (1983) J. Biol. Chem. 259: 12246-12251).

The gene encoding native thrombomodulin has been isolated and sequenced in both its genomic form and as cDNA from several species (Jackman, R., et al., (1986) PNAS 83: 8834-8838 and (1987) 84: 6425-6429, both of which are incorporated by reference). Based on comparisons with known proteins such as the LDL receptor, functional domains are likely to be present (Wen, D., et al., (1987) Biochemistry 26: 4350-4357). According to one study, the fifth and sixth EGF-like (EGF: epidermal growth factor) domains are thought to have the ability to bind thrombin (Kurosawa, S., et al., (1988) J. Biol. Chem. 263: 5993-5996; according to another study, EGF-like domains 4, 5 and 6 are thought to be sufficient as a cofactor for the thrombin-mediated effect of activating protein C (Zushi, et al., (1989) J. Biol. Chem. 264: 10351-10353). The inhibition of the direct coagulation-promoting effect of thrombin (conversion of fibrinogen to fibrin) was attributed to the glycosaminoglycan substituents on the thrombomodulin molecule. attributed (Bourin, M. C. et al., (1986) Pro. Natl. Acad. Sci. USA 83: 5924-5928). The O-linked glycosylation domain provides possible sites for addition of sulphate-containing sugars of this type.

When thrombomodulin is combined with chondroitinase ABC, an enzyme that specifically cleaves certain sulphate-containing carbohydrates bound via O, such as glycosaminoglycans If thrombomodulin is treated with thrombomodulin, the ability of thrombomodulin to inhibit thrombin-mediated platelet aggregation and the thrombin-mediated conversion of fibrinogen to fibrin, the major matrix component of thrombi, is greatly reduced (Preissner, KT, et al., (1990) J. of Biol. Chem. 265-(9): 4915- 4922).

The anticoagulants currently approved for use in humans are not uniformly effective, and there is a need for more effective compounds (see, for example, Prevention of Venous Thrombosis and Pulmonary Embolism, Consensus Development Conference Statement, NIH, 1986, 6(2): 1-23).

In its native form, thrombomodulin is not suitable for anticoagulant therapy because its amino acid sequence makes it membrane-bound and insoluble without treatment with a detergent. It is present in such low amounts (approximately 300 mg thrombomodulin/human) that purification from autopsy or biopsy samples is not practical.

Soluble thrombomodulin-like molecules have been detected in minute amounts in human plasma and urine. These molecules have a reduced ability to promote the activation of protein C and there is a possibility that they have been at least partially inactivated due to oxidation.

These molecules are thought to be degradation products of the membrane-bound molecule (Ishii, H. and Majerus, P., (1985) J. Clin. Inv. 76: 2178- 2181), but they are present in such small amounts that they were difficult to characterize (-0.8 mg/adult male). Proteolysis fragments of the purified native molecule were prepared using trypsin or elastase (see, Ishii, supra, Kurosawa, et al. (1988) J. Biol. Chem. 263: 5593-5996 and Stearns, et al., (1989) J. Biol. Chem. 264: 3352-3356). some of these fragments retain the ability to promote thrombin-mediated activation of protein C in vitro. There is a need for close compounds that have the anticoagulant properties of thrombomodulin, are soluble in plasma, are resistant to inactivation by the action of oxidants, and can be readily produced in large quantities. The present invention meets these and other needs.

SUMMARY

The present invention provides the use of a purified thrombomodulin analog having an extended in vivo circulating half-life, said analog consisting of a polypeptide having an amino acid sequence substantially corresponding to the six epidermal growth factor-like domains, and wherein the six epidermal growth factor-like domains are modified by chemical treatment with hydrofluoric acid to remove glycosylation in the six EGF-like domains, said analog having approximately the native ability to enhance thrombin-mediated activation of protein C compared to native thrombomodulin and having a reduced ability to inactivate thrombin-mediated conversion of fibrinogen to fibrin compared to native thrombomodulin, for the manufacture of a pharmaceutical composition. Preferably, the analogue is soluble in aqueous solution and/or resistant to oxidation.

It is further preferred that the analogue is modified at the sugar residues of the six epidermis growth factor-like domains. The term "modified" means that the six epidermis growth factor-like domains have an altered glycosylation pattern. This can include an exchange and a complete or partial deletion of the native sugar residues. This modification can be achieved by Amino acid residues recognized by cells as glycosylation sites are deleted. However, the sugars can also be removed chemically, either partially or completely. In another modification, the sugars can be treated with enzymes to remove the sulfate substituents. In a further modification, the entire glycosylation domain can be deleted. It is preferred that the analogues to be used have 80% or less of the ability of native thrombomodulin to inactivate the thrombin-mediated conversion of fibrinogen to fibrin compared to native detergent-solubilized rabbit thrombomodulin with respect to equal units of protein C activation.

More specifically, the TM analogs of the invention, when standardized to have equal activity compared to native detergent-solubilized rabbit thrombomodulin in the standard protein C activation assay, have only 80% or less of the activity of the same amount of native thrombomodulin (mass) in a standard assay measuring thrombin-mediated conversion of fibrinogen to fibrin. A preferred analog of the invention has 50% or less of the activity of the same amount of native thrombomodulin in the fibrin assay. These capabilities are measured using standard assays described herein.

The present invention further provides compositions useful for treating thrombosis in mammals comprising an effective amount of a purified thrombomodulin analog having an extended in vivo circulating half-life, wherein the analog consists of a polypeptide having an amino acid sequence substantially corresponding to the six epidermal growth factor-like domains, and wherein the amino acid sequence of the six epidermal growth factor-like domains is chemically treated with hydrofluoric acid to remove the Glycosylation in the six EGF-like domains is modified, wherein the analogue has approximately the same native ability to enhance thrombin-mediated activation of protein C and a reduced ability to inactivate thrombin-mediated conversion of fibrinogen to fibrin compared to native thrombomodulin. The analogues described above for the method are preferred.

DEFINITIONS

The following terms are defined below for the present invention.

"Glycosylation sites" refer to amino acid residues that are recognized by a eukaryotic cell as binding sites for sugar residues. The amino acids to which sugars are attached are typically the residues Asn (N-linkage), threonine, or serine (O-linkage). The specific binding site is signaled by an amino acid sequence, namely Asn-X-(Thr or Ser) for the N-linkage and (Thr or Ser)-X-X-Pro for the O-linkage, where X is any amino acid. The recognition sequence for glycosaminoglycans (a specific type of sulphate-containing sugar) is: Ser-Gly-X-Gly. The terms N-linked and O-linked refer to the chemical group that serves as the binding site between the sugar molecule and the amino acid residue. N-linked sugars are linked via an amino group; O-linked sugars are linked via a hydroxyl group.

"In vivo circulating half-life" refers to the average time required for a mammal to clear one-half of the compound administered to it.

"Native thrombomodulin" refers to the full-length protein as it occurs in nature. If biological activities are determined with respect to the native TM described, this term also includes an aqueous form solubilized with detergent.

"O-linked glycosylation domain" refers to amino acid sequence 463 to 485 of the native thrombomodulin sequence as shown in Table 1.

"Oxidation-resistant analogues" refers to thrombomodulin analogues that are capable of retaining a substantial part of their biological activity after contact with oxidants such as oxygen radicals, chloramine T or hydrogen peroxide.

"Pharmaceutical precursors" refer to nontoxic medically acceptable substances used to complete a medical therapeutic. These substances may be inert, such as water and salt, or they may be biologically active, such as an antibiotic or an analgesic.

"Reduced ability" refers to a statistically significant reduction in a biological trait. The trait is not subject to any restriction, and the measurement or quantification of the trait is standard.

"Sugar residues" refers to hexose and pentose carbohydrates, including glucosamines and other carbohydrate derivatives and moieties, covalently bound to a protein.

"Sulfate substituents" are sulfur-containing acids on pentose or hexose sugars.

"Thrombin-mediated conversion of fibrinogen to fibrin" refers to the enzymatic activity by which thrombin cleaves the precursor protein fibrinogen, producing a fibrin monomer that subsequently polymerizes to form a blood clot.

"Thrombosis" refers to a pathogenic condition in a mammal characterized by the formation of one or more thrombi that are or may be harmful to the mammal's health.

"Thrombomodulin analogues" refers to proteins with an amino acid sequence identical to that of native thrombomodulin, to insoluble and soluble thrombomodulin peptides or fragments, and to oxidation-resistant TM species, all of which have thrombomodulin-like activity. These compounds also include derivatives and amino acid modifications that do not significantly alter the properties of the protein as a cofactor in the activation of protein C compared to native TM.

"Transfer vector" refers to a vector that is co-transfected with a wild-type baculovirus in an insect cell. The transfer vector is designed to promote recombination between the baculovirus genome and the transfer vector, thereby replacing the baculovirus polyhedral gene with a heterologous target gene. If a vector is maintained in a host cell, the vector can either be stably replicated by the cells during mitosis as an autonomous structure or be incorporated into the host genome.

PRECISE DESCRIPTION

According to the present invention, there are provided novel uses and compositions that treat thrombosis with a thrombomodulin(TM) analogue having an increased in vivo circulating half-life, said analogue consisting of a polypeptide having an amino acid sequence substantially corresponding to the six epidermal growth factor-like domains, and wherein the six epidermal growth factor-like domains are modified by chemical treatment with hydrofluoric acid to remove glycosylation in the six EGF-like domains, said analogue having approximately the same native ability to enhance thrombin-mediated activation of protein C and a reduced ability to inhibit thrombin-mediated conversion of fibrinogen to to inactivate fibrin. Pharmacists prefer drug agents that have a specific, limited effect on the patient. Such drug agents are preferred because the risk of causing undesirable side effects is lower than with drug agents that have multiple pharmacological effects on a patient. The present invention has the advantage of treating a single aspect of the coagulation cascade, but does not affect other aspects of the cascade. In another embodiment, the present invention provides uses for increasing the in vivo half-life of TM analogs by modifying the sugar residues of the six epidermal growth factor-like domains at positions 227-462. An increased half-life is advantageous for TM therapy because it allows the administration of smaller amounts of TM compared to the native drug substance while achieving an equivalent pharmacological effect and because a half-life of at least more than a few minutes provides for a more predictable therapeutic treatment.

In addition, these soluble thrombomodulin analogues are economically manufactured and easily purified and administered. Various therapeutic uses are believed to exist, particularly in relation to anticoagulant therapy or thrombosis therapy. In order to fully appreciate the invention, it will now be described in detail.

I. Biological efficacy of thrombomodulin

The pathology underlying thrombosis is that a clot forms in response to a stimulus such as a damaged vessel wall. This stimulus triggers the coagulation cascade and generates thrombin, which has the ability to convert fibrinogen into fibrin, the matrix of the clot. Thrombomodulin is an endothelial cell membrane protein that acts as a receptor for thrombin. In humans, it is distributed on the endothelium of the blood and lymph vessels of all organs except the central nervous system. Thrombin has the ability to bind reversibly to thrombomodulin. When thrombin is bound to thrombomodulin, it is converted from a pro-coagulant enzyme to an anti-coagulant enzyme. The thrombin/thrombomodulin complex inhibits the coagulation cascade in at least two different ways. First, the binding of thrombin to thrombomodulin enhances the thrombin-mediated activation of protein C. Activated protein C inactivates other pro-coagulant components of the coagulation cascade, such as factor Va and factor VIIIa, which in turn inhibits the conversion of further prothrombin to thrombin. Thrombin-mediated activation of protein C is greatly enhanced when thrombin is bound to thrombomodulin, that is, the rate of protein C activation increases at least 1,000-fold. Second, binding to thrombomodulin has direct anticoagulant effects, such as inhibition of thrombin-mediated conversion of fibrinogen to fibrin and thrombin-mediated platelet activation and aggregation. Although thrombomodulin is normally an integral component of the endothelial cell membrane, in the presence of sufficient amounts of detergent it can be detached from the membrane and retains the ability to bind to thrombin in solution.

The preferred thrombomodulin analogues of the present invention protect against thrombus formation when administered systemically because they inhibit the formation of thrombin without disturbing other coagulation parameters such as platelet activation and aggregation. The use of soluble thrombomodulin analogues is therefore effective in preventing thrombus formation, but is safer than native thrombomodulin and other known antithrombotics.

Diseases in which thrombus formation plays a major etiological role include myocardial infarction, disseminated intravascular coagulation, phlebothrombosis, pulmonary embolism, septic shock, acute respiratory distress syndrome, unstable angina and other arterial and venous occlusive diseases. The thrombomodulin analogues of the invention are useful for all of these diseases as well as for other diseases in which pathological thrombus formation is present. By useful is meant that the compounds are useful for treatment, either to prevent the disease or to prevent its progression to a more severe stage. The compounds of the invention also represent a safe and effective anticoagulant agent, for example for patients undergoing bioprostheses such as heart valves. These compounds can be used in the treatment of e.g. B. Pulmonary embolism and acute myocardial infarction replace heparin and warfarin.

In particular, these compounds can be used in the prevention of phlebothrombosis, for example postoperatively. The formation of blood clots in the leg itself is not life-threatening, but is closely associated with the development of pulmonary embolism, which is difficult to diagnose and can be life-threatening. Although various prophylactic treatments are being researched and used clinically, phlebothrombosis and the resulting pulmonary embolism remain a significant problem in many patient groups, particularly in patients undergoing orthopedic surgery. Existing prophylactic treatments such as heparin, warfarin and dextran typically reduce the incidence of phlebothrombosis in patients undergoing orthopedic surgery by over 50% in patients at risk without prophylaxis to 25-30% in treated patients. Serious side effects exist, mainly bleeding-related complications. To minimize bleeding while still maintaining some efficacy, daily laboratory tests and dosage changes are required. Given the disadvantages of existing prophylactics, an antithrombotic agent that effectively prevents phlebothrombosis without predisposing the patient to bleeding could have a significant impact on the patient's recovery and well-being.

Angioplasty is a procedure that is often used to recanalize blocked arteries. Although recanalization can occur, an angioplasty procedure always causes severe damage to the endothelial lining of the artery and blood clots often begin to form. Soluble thrombomodulin analogues administered simultaneously during angioplasty prevent this harmful side effect.

Many acute thromboses and embolisms are currently treated with fibrinolytic therapy to remove the thrombus. The main focus of research is acute myocardial infarction (heart attack). The agents currently used in the treatment of acute myocardial infarction include streptokinase, tissue plasminogen activator and urokinase. The use of these agents can lead to serious bleeding-related complications. In patients in whom a thrombus has been removed by fibrinolytic therapy and blood circulation has been restored, a new blockage of the affected vessel often occurs, ie a new clot forms. Attempts have been made to prevent the new blockages by increasing the dosage or treatment time with a thrombolytic agent. but bleeding is more common in such a case. The therapeutic index of these drugs is therefore low.

The use of thrombomodulin analogues offers protection against the recurrence of occlusion and because they have a local effect, i.e. where thrombin is formed or released from a clot. Therefore, when used in combination with a thrombolytic agent, the dosage of which can subsequently be reduced, the risk of bleeding can be significantly reduced.

The thrombomodulin analogues would be administered by intravenous bolus injection, by continuous intravenous infusion, or by a combination of these two routes. Soluble thrombomodulin mixed with suitable precursors can also be absorbed into the bloodstream from an intramuscular site. Systemic treatment with thrombomodulin analogues can be monitored by determining the activated partial thromboplastin time (APTT) from serially drawn blood samples from the patient. The clotting time observed in this assay is prolonged when sufficient thrombomodulin levels in the bloodstream have been achieved. However, this is a systemic measure of efficacy and it has been discovered by the inventors that an effective dose of a soluble TM analogue does not necessarily affect the APTT. In the present context, a therapeutic effective dose is defined as the amount of a TM analogue required to prevent the formation of pathological clots. Dosage levels and dosing regimens can be adjusted to maintain an appropriate thrombomodulin concentration, measured, for example, by APTT assay.

There are several methods known for the detection and tracking of thrombosis. For example, phlebothrombosis can be diagnosed by contrast venography (Kerrigan, GNW, et al., (1974) British Journal of Hematology 26: 469), Doppler ultrasound (Barnes, RW (1982) Surgery Clinics in North America 62: 489-500), scanning of the uptake of ¹²⁵I-labelled fibrinogen (Kakkar, VV, et al., (1972) Archives of Surgery 104: 156, Kakkar, VV, et al., (1970) Lancet i: 540-542), occlusion plethysmography (Bynum, LJ et al., (1978) Annals of Internal Medicine 89: 162) and thromboscintoscan (Ennis, JT and Elmes, RJ (1977) Radiology 125: 441). These methods are useful in monitoring the effectiveness of the methods and compositions described in this text.

II. TM analogues

A DNA sequence encoding the native human full-length thrombomodulin protein has been isolated (European Patent Application No. 88870079.6, incorporated by reference into the present invention). The cDNA sequence encodes a 60.3 kDa protein of 575 amino acids that includes a signal sequence of approximately 18 amino acids.

The sequences for bovine thrombomodulin, mouse thrombomodulin, and human thrombomodulin show strong homology to each other. Analogous to other proteins, the structure of thrombomodulin can probably be divided into domains. The term "domain" refers to a discrete amino acid sequence that can be associated with a specific function or property. The full-length thrombomodulin gene encodes a peptide precursor with the following domains:

Approximate Amino acid position domain

-18- -1 Signal sequence

1-226 N-terminal domain

227-462 6 EGF-like domains

463-497 O-linked glycosylation

498-521 Stop transfer sequence

522-557 Cytoplasmic domain

See C. S. Yost et al., (1983) Cell, 34: 759-766 and D. Wen et al., (1987) Biochemistry, 26: 4350-4357, both of which are incorporated by reference into the present invention. Compared to native thrombomodulin, the preferred TM analogs of the present invention have been modified to include the six epidermal growth factor-like [EGF] domains with or without the O-linked glycosylation domain. Particularly preferred TM analogues have the following properties: i) they are soluble in aqueous solution without detergents, ii) their activity is retained after contact with oxidants, and iii) when bound to thrombin they enhance thrombin-mediated activation of protein C, but have a reduced ability to inhibit the direct anticoagulant activities of thrombin such as the conversion of fibrinogen to fibrin or the activation and aggregation of platelets. Assays for the latter two assays can be performed using an automated coagulation timer according to the manufacturer's instructions; Medical Laboratory Automation Inc., available from American Scientific Products, McGaw Park, Illinois. (See also H. H. Salem et al., (1984) J. Biol. Chem., 259: 12246-12251, which is incorporated by reference).

In a preferred embodiment, soluble TM analogues are oxidation resistant. This refers refers to analogs that retain activity after contact with oxidants. These analogs are described in detail in co-pending, owned USSN 506,325, filed April 9, 1990, which is incorporated herein by reference.

As used herein, a "soluble TM analogue" means a TM analogue that is soluble in an aqueous solution and capable of being secreted by a cell. For pharmacological administration, the soluble TM analogue or an insoluble analogue with the native cytoplasmic domain may optionally be combined with phospholipid vesicles, detergents or other similar compounds well known to those skilled in the art of galenics. The preferred TM analogues of the invention are soluble in the bloodstream, making the analogues useful for various anticoagulant and other therapies. These modifications have no significant effect on activities of the native thrombomodulin, such as affinity for thrombin or activity in activating protein C.

Two preferred analogs comprise the 6 EGF-like domains; they are 4t/227-462, in which the analog has at least four residues of the human tissue plasminogen activator signal peptide, and 6h/227-462, in which the 6h represents the last six residues of the signal sequence for hypodermin A. More preferably, these analogs are rendered oxidation resistant by substitution of the methionine at position 388 with leucine.

A. General procedures for preparing TM analogues

The present invention encompasses molecular genetic manipulations which can be carried out in a known manner using various methods. The recombinant cells, plasmids and DNA sequences according to the invention represent a means for the production of pharmaceutically useful compounds, whereby the compound secreted by recombinant cells is a soluble thrombomodulin derivative.

In general, the definitions of the terms and descriptions of the general laboratory methods used in this application can be found in J. Sambrook et al., Molecular Cloning, A Laboratory Manual, (1989) Cold Spring Harbor Laboratory, Cold Spring Harbor, New York. This manual is hereinafter referred to as Sambrook and is incorporated by reference into the present application.

All enzymes are used according to the manufacturer’s instructions.

Oligonucleotides that are not commercially available can be prepared according to the solid phase phosphoramidite triester method first described by S. L. Beaucage and M. H. Caruthers, (1981) Tetrahedron Letts., 22 (20): 1859-1862 using an automated synthesizer as described in D. R. Needham-VanDevanter et al., (1984) Nucleic Acids Res., 12: 6159-6168. Oligonucleotides were purified either by native acrylamide gel electrophoresis or by anion exchange HPLC as described in J. D. Pearson and F. E. Regnier, (1983) J. Chrom., 255: 137-149. The size of the nucleotides is given either in kilobases (kB) or in base pairs (bp). These are estimates obtained by agarose or acrylamide gel electrophoresis or from published DNA sequences.

The sequence of the cloned genes and synthetic oligonucleotides can be verified using the chemical degradation method of AM Maxam et al., (1980) Methods in Enzymology, 65: 499-560. The sequence can be confirmed after assembling the oligonucleotide fragments into the double-stranded DNA sequence using the method of Maxam and Gilbert (above) or the chain termination method for sequencing double-stranded templates of RB Wallace et al., (1981) Gene, 16: 21-26. Southern blot hybridization techniques were performed according to Southern et al., (1975) J. Mol. Biol., 98: 503.

In embodiments of the present invention, new peptides and genes must be generated by in vitro mutagenesis. Target genes are isolated in intermediate vectors and cloned into prokaryotes such as E. coli, Bacillus or Streptomyces for amplification. E. coli is most preferred because this organism is easily cultured and better studied than other prokaryotic species. The Sambrook Manual provides enough methods to perform all of the cloning in E. coli described below. Unless otherwise stated, strain MH-1 is preferred. All E. coli strains are grown on Luria broth (LB) with glucose or M9 medium supplemented with glucose and amino acids from a casein acid hydrolysate. Strains with antibiotic resistance were maintained at the drug concentrations specified by Sambrook. Transformations were performed according to the method described by D. A. Morrison, (1977) J. Bact., 132: 349-351 or by J. E. Clark-Curtiss and R. Curtiss, (1983) Methods in Enzymology, 101: 347-362, Eds. R. Wu et al., Academic Press, New York. Typical vector cells include pBR322 and the pUC series, which are commercially available.

B. Gene synthesis

The gene encoding native thrombomodulin has been isolated and sequenced from several species both in its genomic form and as cDNA (R. Jackman, et al., (1986) PNAS 83: 8834-8838 and (1987) 84: 6425-6429, both of which are incorporated by reference).

The publication of the full-length DNA sequence encoding human thrombomodulin and thrombin facilitates the engineering of genes and will be used as a starting point for the construction of DNA sequences encoding TM peptides. The peptides of the invention are preferably soluble derivatives which lack the stop transfer sequence of TM and which also have internal amino acid substitutions. In addition, these analogues are secreted by eukaryotic cells which have been transfected or transformed with plasmids containing genes encoding these polypeptides. The methods for making modifications such as amino acid substitutions, deletions or the attachment of signal sequences to cloned genes are known. Specific methods used here are described below.

The full-length thrombomodulin gene can be prepared by several methods. Human genomic libraries are commercially available. Oligonucleotide probes specific for these genes can be synthesized using the published gene sequence. Methods for screening genomic libraries with oligonucleotide probes are known. The publication of the thrombomodulin gene sequence shows that no introns exist within the coding region. A genomic clone therefore provides the necessary starting material for preparing a thrombomodulin expression plasmid using known methods.

A DNA fragment encoding thrombomodulin can be found by exploiting restriction endonuclease cleavage sites identified in the flanking regions or existing within the gene (R. W. Jackman et al., (1987) Proc. Natl. Acad. Sci. USA., 84: 6425-6429).

However, the full-length genes can also be obtained from a cDNA library. For example, messenger RNA prepared from endothelial cells is a suitable starting material for the preparation of cDNA. A cDNA molecule containing the gene encoding thrombomodulin is identified as described above. Methods for preparing cDNA libraries are well known (see Sambrook above).

Genes encoding TM peptides can be made from wild-type TM genes that were first constructed using the gene encoding full-length thrombomodulin. A preferred method for producing wild-type TM peptide genes for subsequent mutation combines the use of synthetic oligonucleotide primers with polymerase extension on an mRNA or DNA template. This polymerase chain reaction (PCR) method amplifies the desired nucleotide sequence. This method is described in U.S. Patents 4,683,195 and 4,683,202. Cleavage sites for restriction endonucleases can be incorporated into the primers. Genes amplified by PCR can be purified from agarose gels and cloned into an appropriate vector. Changes in the natural gene sequence can be introduced by in vitro mutagenesis techniques or by using the polymerase chain reaction with primers designed to contain appropriate mutations.

The TM peptides described herein are secreted when expressed in cultured eukaryotic cells. Secretion can be achieved by using the native signal sequence of the thrombomodulin gene. However, genes encoding the TM peptides of the present invention can also be ligated in frame to a signal sequence other than the signal sequence corresponding to the native thrombomodulin gene. For example, the signal sequence of t-PA (see WO 89/00605, incorporated by reference) or of hypodermin A or B (see EP 326,419, incorporated by reference) can be linked to the polypeptide (see Table 2). In the preferred embodiment of the present invention, the signal sequence of t-PA containing the second intron of the human t-PA gene is used.

The inclusion of the intron increases the productivity of the neighboring structural gene.

In the analogs of the present invention, those portions of the gene encoding the stop transfer and cytoplasmic domains of the carboxy-terminal region of the native ihrombomodulin gene are deleted. A stop codon must therefore be added to terminate translation at the desired site. However, a stop codon can also be provided by the desired expression plasmid. In addition, a polyadenylation sequence is required to ensure that the mRNA is properly processed in the eukaryotic cells encoding the TM analog. In addition, an initiation codon for expression of the TM peptides may need to be provided if none is present. Such sequences can be provided by the native gene or by the expression plasmid.

Cloning vectors suitable for replication and integration in prokaryotes or eukaryotes and containing transcription and translation stops, start sequences and promoters suitable for regulating expression in TM peptides are described in the present text. The vectors consist of expression cassettes containing at least one independent stop sequence, sequences allowing replication of the plasmid in both eukaryotes and prokaryotes, i.e. shuttle vectors, and selection markers for both prokaryotic and eukaryotic systems.

C. Expression of TM peptides in prokaryotic cells

In addition to using cloning methods in E. coli to amplify the cloned sequences, it may be desirable to express TM analogues in prokaryotes. As discussed in more detail below, the carbohydrate portions of the mature protein are essential for activity as a cofactor for the Activation of protein C is not essential, but does have an impact on the direct anticoagulant properties of the TM analogues as well as on the circulating half-life of the molecule. Expression of thrombomodulin analogues in E. coli has provided a useful tool to investigate this aspect. E. coli transformed with an expression plasmid encoding a soluble TM analogue can yield a therapeutically functional protein.

Methods for expressing cloned genes in bacteria are well known. To obtain a high level of expression of a cloned gene in a prokaryotic system, it is essential to construct expression vectors containing at least one strong promoter that controls the termination of mRNA transcription. Regulatory regions suitable for this purpose include, for example, the promoter and operator region of the E. coli β-galactosidase gene, the E. coli tryptophan biosynthesis pathway, or the left promoter of phage lambda. The incorporation of selection markers into DNA vectors that are transformed into E. coli is useful. Such markers include, for example, the genes that specify resistance to ampicillin, tetracycline, or chloramphenicol.

For details regarding selectable markers and promoters for use in E. coli, see Sambrook. In the described embodiment of the present invention, pUC19 is used as a vector for subcloning and amplifying the desired gene sequences.

D. Expression of TM peptides in eukaryotic cells

It is assumed that the person skilled in the art is aware of the expression systems chosen for the expression of the desired TM peptides and no attempt will be made to describe the various known methods to accurately describe the expression of proteins in eukaryotes.

The DNA sequence encoding a soluble TM analogue can be ligated to various expression vectors used to transform host cell cultures. Typically, the vectors contain marker genes and gene sequences for initiating transcription and translation of the foreign gene.

Preferably, the vectors contain a marker gene that provides a phenotypic trait for the selection of transformed host cells, such as dihydrofolate reductase, metallothionein, hygromycin or neomycin phosphotransferase. The core polyhedron virus protein from Autographa californica is useful for screening transfected Spodoptera frugiperda and Bombyx mori insect cell lines to identify recombinants. Known selectable markers for yeast are Leu-2, Ura-3, Trp-1 and His-3 (Gene (1979) 8: 17-24). There are numerous other markers, both known and unknown, that comply with the above-mentioned scientific principles, and all of them would be suitable as markers to detect those eukaryotic cells that have been transfected with the vectors according to the invention.

Among the higher eukaryotic cell systems suitable for the expression of TM analogues, several cell systems are available. Illustrative examples of mammalian cell lines include RPMI 7932, VERO, HeLa cells, Chinese hamster ovary (CHO) cell lines, WI38, BHK, COS-7, C127 or MDCK cell lines. CHL-1 is a preferred mammalian cell line. When CHL-1 is used, hygromycin is added as a eukaryotic selection marker. CHL-1 cells are derived from RPMI 7932 melanoma cells, a readily available human cell line. The CHL-1 cell line was deposited with the ATCC under the terms of the Budapest Treaty on June 18, 1987 under the number CRL 9446. Cells suitable for this invention are available for purchase from the American Type Culture Collection. Illustrative insect cell lines include Spodoptera frugiperda (army armyworm) and Bombyx mori (silk moth).

As indicated above, the expression vector, e.g. the plasmid, used to transform the host cell preferably contains gene sequences for transcription initiation as well as control sequences for translation of the TM peptide gene sequence. These sequences are referred to as expression control sequences. If the host cell is derived from an insect or mammal, illustrative expression control sequences include, but are not limited to, the retrovirus long terminal repeat promoter ((1982) Nature, 297: 479-483), the SV40 promoter ((1983) Science, 222: 524-527), the thymidine kinase promoter (J. Banerji et al., (1982) Cell, 27: 299-308), or the beta-globin promoter (P. A. Luciw et al., (1983) Cell, 33: 705-716). The recipient vector nucleic acid containing the expression control sequences is cleaved with restriction enzymes and sized as needed or desired. This segment is ligated to a DNA sequence encoding the TM peptide using well-known methods.

If host cells from higher animals are used, polyadenylation or transcription termination sequences must be incorporated into the vector. An example of a polyadenylation sequence is the SV40 polyadenylation sequence, which can also serve as a transcription terminator.

The genes incorporated into the respective vectors can be used to control protein synthesis either in transient expression systems or in stable clones. In the first case, yields are low, but the experiments do not take long. In the latter case, the isolation of high-producing clones is more time-consuming. For the different types of experiments Different vectors can be used. In particular, in transient expression, sequences can be introduced into the plasmid that allow the plasmid to replicate in high copy numbers within the cell. These sequences can originate from a virus such as SV40 (e.g. BC Doyle et al., (1985) J. Cell Biol., 100: 704-714) or from chromosomal replication sequences such as the autonomous replication sequences from the mouse (Weidle et al., (1988) Gene, 73: 427-437). The vector for transient expression should therefore contain a strong promoter such as the early SV40 promoter (e.g. BA von Zonnenfeld et al., (1987) Proc. Natl. Acad. Sci. USA., 83: 4670-4674) to direct transcription of the gene of interest. Although transient expression is a rapid method for testing gene products, the plasmid DNA is not incorporated into the host cell chromosome. Therefore, the use of transient expression vectors does not result in stably transfected cell lines. A description of a plasmid suitable for transient expression can be found in A. Aruffo & B. Seed, (1987) Proc. Natl. Acad. Sci. USA., 84: 8573-8577.

However, TM analogs can also be produced in the insect cell lines described above using the baculovirus system. This system was described by VA Luckow and MD Summers (1988) Bio/Technology, 6: 47-55. In general, this expression system allows a higher level of expression than most mammalian systems. The baculovirus infects the host insect cell, replicates its genome in numerous cycles, and then produces large quantities of polyhedral crystals. The polyhedral gene can be replaced by a TM peptide gene. The polyhedral promoter then makes large quantities of the analog protein following infection of the culture host cell and replication of the baculovirus genome. The non-secreted gene product is harvested from the host 3-7 days after infection. The TM peptide can However, if corresponding signal sequences are present on the protein, they can also be secreted from the cells.

The host cells are competent or are made competent for transfection in various ways. Several well-known methods exist for introducing DNA into animal cells. These include: calcium phosphate precipitation, the DEAE-dextran technique, fusion of the recipient cells with bacterial protoplasts containing the DNA, treatment of the recipient cells with liposomes containing the DNA, electroporation, and microinjection of the DNA directly into the cells. See B. Perbal, "Practical Guide to Molecular Cloning," 2nd ed., John Wiley & Sons, New York, and Wigler, et al., (1987) Cell, 16: 777-785.

E. Cell culture

Preferably, the host cell can be rapidly grown in cell culture and is capable of appropriately glycosylating the expressed gene products. Cells known to be suitable for dense growth in tissue culture are particularly desirable and various invertebrate or vertebrate cells, both normal and transformed cell lines, have been used in this area.

The transfected cells are grown using methods known in the art. See, for example, Biochemical Methods in Cell Culture and Virology, Kuchler, R. J., Dowden, Hutchinson and Ross, Inc. (1977). In systems in which the protein is secreted from the host cell, the expression products are harvested from the cell medium or from the cell suspension after disrupting the host cell system, e.g. by mechanical or enzymatic means, as is well known to those skilled in the art.

F. Purification of TM analogues

The TM peptides according to the present invention are preferably secreted by recombinant eukaryotic cells in culture. The TM analogues are produced in serum-free medium or in a medium with serum supplement and secreted intact. If prokaryotic cells are used, the TM analogues can be secreted intracellularly. The peptides can be fully or partially glycosylated or non-glycosylated. After growth of the recombinant cells and simultaneous secretion of TM analogues into the culture medium, this "conditioned medium" is harvested. The conditioned medium is then clarified by centrifugation or filtration to remove cells and cell debris. The proteins contained in the clarified medium are concentrated by adsorption to any suitable resin such as Q-Sepharose or metal chelators, or by ammonium sulfate fractionation, polyethylene glycol precipitation, or by ultrafiltration. Other methods known in the art may also be suitable. The TM analogues may be further purified as described in Galvin, J. B., et al., (1987) J. Biol. Chem., 262: 2199-2205 and Salem, H. H. et al., (1984) J. Biol. Chem., 259: 12246-12251, and in the manner described in the present embodiment. Where appropriate, additional protein purification techniques such as affinity chromatography, ion exchange chromatography, size separation chromatography or other protein purification techniques may be required for the purification of TM analogues secreted by cells in culture.

The recombinant TM analogues can be produced in multiple conformations that are detectable under non-reducing chromatography conditions. It is desirable to remove those species that have low specific activity, which is achieved by various chromatographic techniques, including anion exchange or size separation chromatography.

The recombinant TM analogues can be concentrated by pressure dialysis and buffers that are directly exchanged for volatile buffers (e.g. N-ethylmorpholine (NEM), ammonium bicarbonate, ammonium acetate and pyridine acetate). In addition, samples can be freeze-dried directly from these volatile buffers, yielding a stable protein powder free of salts and detergents. In addition, freeze-dried samples of recombinant analogues can be effectively resolubilized in buffers suitable for infusion (e.g. phosphate buffered saline) before use. Other suitable buffers include, for example, hydrochloride, hydrobromide, sulfate acetate, benzoate, malate, citrate, glycine, glutamate and aspartate.

G. Oxidation-resistant TM analogues

Native thrombomodulin is susceptible to oxidation and, when oxidized, loses its ability to promote the activation of protein C. Many conditions requiring anticoagulation are also associated with large amounts of toxic oxygen radicals that can inactivate biological molecules and cause significant tissue damage. Such conditions include reperfusion injury associated with myocardial infarction, disseminated intravascular coagulation associated with septicemia, and alveolar fibrosis associated with adult respiratory distress syndrome (see Otani, H., et al., (1984) Circ. Res. 55: 168-175, Saldeen, T., (1983) Surg. Clin. NA 63(2): 285-304, and Idell, S., et al., (1989) J. Clin. Inv. 84: 695-705.) In addition, any wound, such as a surgical wound, results in the infiltration of activated monocytes, polymorphonuclear leukocytes, etc., which can generate toxic oxygen species and release various proteolytic enzymes such as elastase. The connection between endothelial cell damage, inflammation and thrombosis has been recognized for a long time (see Example: The Molecular and Cellular Biology of Wound Repair, eds. Clark, RAF and PM Henson 1988). Thrombomodulin is inactivated upon contact with toxic oxygen species and this is thought to play an essential role in many disease states.

Methods by which amino acids, in particular methionines, can be made oxidation-resistant are well known to those skilled in the art. Thiol groups can, for example, be chemically modified with iodoacetic acid to form the oxidation-resistant sulfonium (Gundlach, H. G., et al., (1959) J. Biol. Chem. 234: 1754). A preferred method consists in removing the sensitive amino acid or replacing it with one or more different amino acids that do not react with oxidants. Particularly preferred amino acids are the amino acids leucine, alanine and glutamine due to their size and neutral character. Two methionines of thrombomodulin that are susceptible to oxidation are the methionines at residues 291 and 388. If only one methionine is to be blocked or removed, this is preferably the methionine at position 388.

Methods by which amino acids in the sequence of a protein can be removed or replaced are well known. For example, genes encoding a peptide with an altered amino acid sequence can be produced synthetically. A preferred method is in vitro site-directed mutagenesis. Site-directed mutagenesis uses a synthetic oligodeoxyribonucleotide containing a desired nucleotide substitution, insertion or deletion that is such that it specifically alters the nucleotide sequence of a single-stranded target DNA. Hybridization of this oligonucleotide, also called a primer, with the single-stranded matrix and subsequent extension of the primer produces a heteroduplex DNA that, when replicated in a transformed cell, encodes a protein sequence with the desired mutation.

To determine resistance to oxidative loss of thrombomodulin activity, the test material (100-250 ug/ml) is first incubated with an oxidizing agent such as chloramine T or hydrogen peroxide at a concentration of 5-10 mM chloramine T or 200-1000 mM hydrogen peroxide for 20 minutes at room temperature in a buffer of 0.2% N-ethylmorpholine and 0.008% Tween 80 at pH 7.0. Following this contact with the oxidizing agent, the test material is evaluated using one of the bioactivity assays described below, specifically for the ability to act as a cofactor for the activation of protein C. Those TM analogue mutants that still have at least 60%, preferably 90%, of the activity they had before contact with the oxidizing agents are considered oxidation-resistant compared to the wild-type TM analogue (the non-mutated TM analogue) or native thrombomodulin.

H. Laboratory assays for measuring TM activity

There are several laboratory assays for measuring TM activity. The activity of the protein C cofactor can be measured using the assay described by Salem, et al., (1984) J. Biol. Chem. 259(19): 12246-12251 and Galvin, et al., (1987) J. Biol. Chem. 262(5): 2199-2205. Briefly, this assay consists of two steps. The first step is the incubation of the test TM analogue with thrombin and protein C under defined conditions (see examples below). In the second step, the thrombin is inactivated with hirudin or antithrombin III and heparin, and the activity of the newly activated protein C is determined with a chromogenic substrate, whereby the chromophore is activated by the proteolytic activity of the activated Protein C is released. This assay is performed with purified reagents.

However, the effect of a TM analogue can also be measured using plasma in clotting time assays such as activated partial thromboplastin time (APTT), thrombin time (TCT) or prothrombin time (PT). These assays distinguish between different mechanisms of anticoagulation and involve the activation of protein C. A prolongation of the clotting time in any of these assays indicates that the molecule is capable of inhibiting clotting in plasma.

The above-mentioned assays identify soluble TM analogues that are able to bind thrombin and activate protein C both in purified systems and in the plasma environment. Subsequently, further activities of the native thrombomodulin, such as the inhibition of thrombin-catalyzed binding of fibrin from fibrinogen (Jakubowski, et al., (1986) J. Biol. Chem. 261(8): 3876-3882), the inhibition of thrombin activation by factor V (Esmon, et al., (1982) J. Biol. Chem. 257: 7944-7947), the accelerated inhibition of thrombin by antithrombin III and the heparin cofactor II (Esmon, et al., (1983) J. Biol. Chem. 258: 12238-12242), the inhibition of thrombin activation by factor VIII (Polgar, et al., (1987) Thromb. Haemostas. 58: 140), the inhibition of thrombin-mediated inactivation of protein 5 (Thompson and Salem, (1986) J. Clin. Inv. 78(1): 13-17) as well as the inhibition of thrombin-mediated activation and aggregation of platelets (Esmon, et al., (1983) J. Biol. Chem. 258: 12238-12242) were evaluated in further assays.

In the present invention, not all TM analogues have activities equivalent to that of native thrombomodulin. For example, comparing an amount of a TM analogue of the present invention with a corresponding amount of native Thrombomodulin (measured in units of protein C cofactor activity, see definition below), the TM analogue has a reduced ability to inhibit the thrombin-mediated conversion of fibrinogen to fibrin by at least 20%, preferably 50%, compared to native thrombomodulin.

I. Methods for altering the glycosylation of TM analogues

The carbohydrate substituents on proteins can affect both biological activity and circulating half-life. To prepare the TM analogs of the present invention, an O-linked glycosaminoglycan carbohydrate such as that found in the native thrombomodulin protein must be absent. This can be achieved in a number of ways. One method is to treat the protein containing an O-linked carbohydrate with a glycanase known to specifically cleave sulfate-containing glycosaminoglycans, for example, chondroitinase ABC or hyaluronidase. This method is described in the work of Bourin, M, et al., (1988) JBC 263(17): 8044-8052, which is incorporated by reference into the present invention.

A second method for removing the O-linked carbohydrate is to introduce site-directed mutations into the protein. The attachment of glycosaminoglycans is controlled by the consensus amino acid recognition sequence X-serine glycine-X-glycine-X, where X is any amino acid (Bourdon, MA, et al., (1987) PNAS. USA 84: 3194-3198). The recognition sequence for other types of O-linked sugars is threonine/serine-XX-proline. The O-linked domain of thrombomodulin has one potential glycosaminoglycan attachment site (aa 472) and three other potential attachment sites for O-linked carbohydrates (aa 474, 480, and 486). Any A change introduced into the nucleotide sequence which removes or alters the identity of any one or more amino acids in that recognition sequence removes the potential attachment site for O-linked carbohydrates. Methods for introducing site-directed mutations into a nucleotide sequence are described above.

A preferred method of removing O-linked carbohydrates from a TM analogue is to prepare an analogue peptide that does not contain the amino acids considered to be the O-linked domain, i.e., amino acids 468 to 485 of the native thrombomodulin gene sequence shown in Table 1. Methods by which this can be achieved are well known to those skilled in the art and have been described above.

The circulating half-life of a protein can be altered by the amount and composition of carbohydrates attached to it. The TM analogs of the present invention contain both O-linked and N-linked carbohydrates. In addition to the potential glycosylation sites discussed above, potential N-linked sites exist at amino acids 364, 391, and 393, and potential O-linked sites exist at amino acids 319, 393, and 396. In addition to the methods described above, other methods for altering carbohydrate composition include: 1) expression of the TM analog gene in bacteria such as E. coli, which do not have the cellular machinery required for glycosylation of mammalian proteins, 2) expression of the TM analog gene in various eukaryotic cells, since each has its own characteristic enzymes responsible for the attachment of the characteristic sugar residues, and 3) treatment with chemicals such as hydrofluoric acid. For example, hydrofluoric acid chemically degrades sugars with acidic and neutral pH, while it chemically degrades basic sugars such as N-acetylglucosamine and, under certain conditions, galactosamine.

J. Formulation and use of thrombomodulin analogues

The soluble TM analogues described in this text may be prepared as a lyophilised or liquid formulation. The material shall be provided in a concentration suitable for pharmaceutical use as either an injection or infusion preparation.

These compounds can be administered alone or as mixtures with other physiologically acceptable active substances such as antibiotics, other anticoagulants, single-chain t-PA or inert substances or with suitable carriers such as water or normal saline. The analogues can be administered parenterally, for example by injection. The injection can be subcutaneous, intravenous or intramuscular.

These compounds are administered in pharmaceutically effective amounts and often as pharmaceutically acceptable salts such as acid addition salts. Such salts may include, for example, the hydrochloride, hydrobromide, phosphate, sulfate, acetate, benzoate, malate, citrate, glycine, glutamate and aspartate, among others. The analogues described in the present text may exhibit enhanced in vivo activity by incorporation into micelles. Methods for incorporation into micelles of ionic detergents or phospholipid micelles are known.

An antithrombotic agent may be prepared using the soluble TM analogs described herein and may consist solely of a fully purified analog or of a fully purified analog in combination with a thrombolytic agent as described above. Compounds of the present invention which have been shown to have the above-mentioned physiological effects can be used for numerous therapeutic purposes such as inhibition of blood coagulation. These compounds can therefore be used as therapeutics for the treatment of various circulatory disorders such as coronary embolism, pulmonary embolism, strokes, as well as in reocclusion prophylaxis following thrombolytic therapy, and these compounds are suitable for preventing the progressive enlargement of a clot during a heart attack. The disclosed compounds can also be useful for the treatment of systemic coagulation disorders such as disseminated intravascular coagulation (DIC), which is frequently associated with blood poisoning, certain carcinomas and pregnancy toxemia.

These compounds can be administered to mammals, both for veterinary purposes as in domestic animals and for clinical use in humans in a manner similar to other therapeutic agents, i.e. in a physiologically acceptable carrier. Generally, the dosage of TM analogue administered is approximately 0.0002 to 5000 µg/kg, more usually 0.02 to 500 µg/kg of host body weight. These dosages can be administered as a continuous infusion over a prolonged period until a desired circulating level is achieved or, preferably, as a bolus injection.

EXAMPLES EXAMPLE 1. Isolation and expression of TM analogue sequences A. Cloning

The genes for the production of recombinant thrombomodulin analogue peptides were published as in the co-pending applications on April 28, 1989 under US Serial No. 345,372, on September 13, 1989 under US Serial No. 406,941 and on February 16, 1990 under PCT- Serial No. 90/00955, which are incorporated herein by reference. Briefly, a gene encoding the 6 EGF-like domains of thrombomodulin corresponding to amino acids 227-462, as well as other portions of the thrombomodulin peptide, was isolated using human DNA (see Table 1). This DNA was isolated from fetal liver by the method of Blin, N and DW Stafford, (1976) Nucleic Acids Res. 3: 2302. The DNA was then used as a template in a polymerase chain reaction with synthetically prepared primers selected to encompass the desired regions (see Tables 3 & 4).

i. Isolation of the genes encoding amino acids 227-462

The following steps provide a route to obtain a DNA insert encoding amino acids (aa) 227-462 and uses for primers #1033 and #1034 (see Table 3). Of course, by modifying the procedures below using other primers, other soluble TM analogues can be obtained.

The sequence of primers #1033 and #1034 corresponds to the 5' and 3' termini of the desired domain, but they were modified to contain a BamHI site. A termination codon (TGA) was introduced following base 1586. The polymerase chain reaction was carried out under the conditions described by Saiki, et al., (1988) Science 320: 1350-1354, except that the initial annealing temperature was 37ºC. After 10 cycles, the annealing temperature was increased to 45ºC for the remaining 30 cycles. An aliquot of the reaction products was separated on a 5% polyacrylamide gel and visualized by staining with ethidium bromide. A band of the predicted size (700 bp) was clearly visible. To confirm the identity of this band, it can either be sequenced or hybridized with an internal probe.

ii. Isolation of genes encoding other thrombomodulin regions

The polymerase chain reaction as described here was used in a similar manner to isolate additional thrombomodulin fragments corresponding to the regions listed in Table 4. Specifically, these regions include one or more of the EGF-like domains as well as the O-linked glycosylation domain. The sequences of the primers selected to generate the desired regions are shown in Table 3.

iii. Cloning plasmids containing thrombomodulin analogue genes

The remainder of the polymerase chain reaction mixture (i) described above was digested with BamHI and separated on a 5% polyacrylamide gel, and the 700 bp band was excised and eluted. It was ligated with pUC19 digested with BamHI and the new plasmid was transformed into E. coli strain DH5-alpha. Recombinant colonies were selected on medium containing ampicillin and 5-bromo-4-chloro-3-indolyl-β-D-galactoside. White colonies were separated on a grid and hybridized using the Grunstein-Hogness technique with a synthetically obtained gene corresponding to aa 283-352 of thrombomodulin, which was excised from a cloning plasmid (pTM2.1) with EcoRI and HindIII and then labeled with ³²p by random priming (Boehringer Mannheim).

After treating an X-ray film with the filters, the one colony that hybridized with the pTM2.1 probe was selected and a culture was grown. The DNA was extracted and analyzed by restriction with BamHI or BglII to confirm the presence of an insert with the correct restriction map. The excised insert was also transferred to nitrocellulose and analyzed by hybridization with labeled pTM2.1.

By both methods, it was confirmed that the 700 bp insert contained the coding sequence for the 6 EGF-like domains of thrombomodulin. The insert was sequenced to verify that no mutations were inadvertently introduced during PCR. The plasmid containing the desired gene fragment is designated pUC19pcrTM7.

B. Expression of TM 1. Construction of the AcNPV transfer vectors

The transfer vectors described below are also described in copending application U.S. Serial No. 345,372. The transfer vectors contain the hypodermin A signal sequence from Hypoderma lineatum.

i. pHY1 and pSC716.

The oligomers containing the hypodermin A signal sequence, a translation initiation codon, a BglII restriction site, a 5' BamHI protruding end, and a 3' Kpn1 protruding end, COD #1198 and COD #1199 (see Table 2) were allowed to anneal and cloned into pSC654, a pUC19 derivative, to generate pHY1. Plasmid pHY1 was digested with BamHI and EcoRI, releasing the hypodermin A signal sequence. This sequence was then ligated into pSC714 to yield vector pSC716. Plasmid pSC714 is a derivative of pVL1393, which was described by Summers, et al. was obtained. The only difference between the two is that in pSC714 one of the BglII cleavage sites had been destroyed.

ii. Construction of pHY101

The BamHI fragment from pUC19pcrTM7 (see Aiii above) was cloned into the BglII site of pHY1, and the orientation was chosen so that the hypodermin A signal sequence was adjacent to amino acid 227. This plasmid was pHY101.

iii. Construction of the AcNPV transfer vector pTMHY101.

Plasmid pHY101 was treated with BamHI/EcoRI, which combination releases the hypodermin A signal sequence linked to the coding sequence for the TM analogue. The shuttle vector pVL139B contains a partially deleted AcNPV polyhedrin gene and a single BamHI or EcoRI restriction site. The BamHI/EcoRI fragment from pHY101 was inserted downstream of the polyhedrin promoter, yielding a plasmid pTMHY101 in which the hybrid gene was under the control of the polyhedrin promoter.

iv. Construction of the other AcNPV transfer vectors.

Transfer plasmids containing other TM analogue gene sequences were constructed similarly as outlined above. Fragments of the cloning plasmids described above were cloned in frame into pSC716 so that the TM analogue gene sequence was fused to the hypodermin A signal sequence. The TM gene sequences are listed in Table 4.

v. Production of pure phage stocks

Cells were transfected using a calcium phosphate precipitation technique modified for insect cells by Summers and Smith (1974). Briefly, a T25 flask was seeded with 2 x 106 Sf9 cells and the cells were allowed to attach for 1 hour at room temperature. Two µg of transfer vector, e.g. pTHR28, and 1 µg of AcNPV DNA were coprecipitated in calcium phosphate and incubated with the cells for 4 hours. The cells were rinsed and refilled with growth medium and then placed in an incubator at 28°C for 3-4 days. During this incubation, the cells produce both recombinant and non-recombinant viruses that accumulate in the growth medium. This medium, containing a mixed virus stock, was assayed for the Presence of protein C cofactor activity checked (see below).

Recombinant viruses were detected using a plaque assay. Transfection stocks were diluted (10-4, 10-5, and 10-6) and plated 4-7 days after transfection. Occlusion-negative (recombinant) plaques were singled out and replated (at a dilution of 10-1, 10-2, and 10-3) 7 days after plating. After an additional 7 days, plates contained 100% pure occlusion-negative recombinant plaques. A single PFU of each was selected for production. A high titer virus stock was grown by infecting 5 ml of Sf9 cells (1 x 106/ml in Excell 400 medium (JR Scientific)) with a single PFU and growing for 4-5 days. A portion of this stock was then diluted 1:50 - 1:100 with Sf9 cells grown to mid-10g phase, creating a protein stock.

2. Production of human TM analogues in mammalian cells i. Mammalian expression vectors for the TM analogues

In this example, a mammalian expression vector is provided containing the analog genes of Example 1A. The genes are operably linked to the human tissue plasminogen activator signal sequence (see Table 2). The expression plasmid, pPa124, contains a promoter for expression of the cloned genes contained within three copies of the long terminal repetitive sequences derived from Harvey sarcoma virus. This plasmid was derived from pPA119 and pSC672, both of which are described in detail in copending U.S. application Serial No. 251,159, filed September 29, 1988, which is incorporated by reference. A BglII- BclI fragment encoding the SV40 polyadenylation region was isolated from pSC672. This fragment was cloned into pPA119 which had been digested with BglII and BclI. In the resulting plasmid pPA124, both the BglII and BclI cleavage sites remained intact. Plasmid pPA124 contains the t-PA signal sequence next to a suitable restriction site, and this signal sequence also contains the second intron of the human t-PA gene.

The gene encoding the soluble TM analogue was removed from pUC19pcrTM7 by treatment with BamHI and ligated to BglII-treated pPA124. The transformants were screened for the presence of the insert in the correct orientation, that is, in the orientation in which the t-PA signal sequence was linked to the 5' end of the thrombomodulin insert encoding an open reading frame. This plasmid pTM101 was then digested with ClaI and ligated to a ClaI fragment containing the dhfr gene under the control of the SV40 promoter. The ClaI fragment is described in WO88/02411 on page 26. The transformants were screened for the presence of this dhfr cassette and then the orientation relative to the plasmid was determined by restriction mapping (pTM103).

Plasmid pTM103, which contains the dhfr sequence in the opposite orientation to the thrombomodulin sequence, was treated with BclI and a DNA fragment encoding a gene providing hygromycin resistance on a BamHI fragment was ligated into the plasmid. After transformation into E. coli strain DH5α, clones were selected for their ability to grow on plates containing both ampicillin and hygromycin B. The orientation of the hygromycin B gene with respect to the plasmid was determined by restriction mapping. A plasmid, pTM108, in which the hygromycin B gene is in the opposite orientation to the TM gene was grown in culture. In this plasmid, the sequences encoding the TM analogue encoded under the control of the triple LTR promoter, with both a gene conferring hygromycin B resistance and one encoding dhfr on the plasmid. A similar expression plasmid, pTHR13, also contains the t-PA signal sequence operably linked to the sequence encoding the 6 EGF-like domains, both under the control of the cytomegalovirus promoter. This plasmid contains the M13 origin of replication, making it suitable for the in vitro site-directed mutagenesis described below. The thrombomodulin sequence has been linked to the tissue plasminogen activator signal sequence, ensuring its secretion. The TM analogue produced by these two plasmids, namely 4t/227-462, consists of the 6 EGF-like domains of thrombomodulin with 4 additional amino acids at the N-terminal end, which represent the result of processing of the t-PA signal peptide.

ii. Transfection, selection and amplification of stable mammalian clones.

For transfection, 10 µg of pTM108 was mixed with Lipofectin reagent (Bethesda Research Laboratories) and added to a monolayer of 10⁵ CHL-1 host cells in 6-well plates. Forty-eight hours after transfection, a known number of cells were plated on selective media. Resistance to hygromycin B was used as a selection marker. CHL-1 cells transfected with the bacterial hygromycin B gene can still grow on 0.3 mg/ml hygromycin B.

The transfection or selection frequency was 2/10³ and was determined as the number of colonies that formed after selection divided by the total number of cells plated. The culture supernatant was shown to contain 1.5 U/ml TM activity after 24 hours in contact with the cells.

A population of cells resistant to the first selection conditions was then subjected to a second round of selection pressure. The growth medium was supplemented with 100 nM or 500 nM methotrexate (MTX) to select transfectants expressing the dhfr gene. Only clones that had amplified the dhfr gene would be able to grow at this high MTX concentration. During gene amplification, other plasmid sequences are co-amplified along with the dhfr gene and therefore also lead to increased gene expression of the non-selectable gene. Resistant clones appeared after 5 to 6 weeks. Individual clones that were resistant to these MTX concentrations were isolated and assayed. A culture selected in 100 nM MTX was shown to produce 4.9-14.7 U of protein C activating activity per ml (see below). A pooled population was plated at a tenfold higher MTX concentration (1 uM and 5 uM, respectively). Again, clones were recovered from this selection step and assayed. Clones were shown to produce TM analogue at each step and secrete it into the culture medium.

C. Site-directed mutagenesis

The region of native thrombomodulin containing the 6 EGF-like domains contains two methionine residues, one at position 291 and one at position 388 (see Table 1). In vitro site-directed mutagenesis was used to convert one or both of these methionines into other amino acids. In site-directed mutagenesis, a synthetic DNA sequence containing a desired nucleotide substitution, insertion or deletion is used to specifically alter the nucleotide sequence of a single-stranded template DNA. Hybridization of this synthetic DNA with the template and subsequent extension of the primer produces a heteroduplex DNA that can be used to target cells. to obtain the desired mutation. A scheme illustrating this process is shown in Fig. 1.

A plasmid for producing single-stranded DNA copies, pTHR14, was constructed by ligating the F1 origin of replication contained on an AseI-ScaI fragment into an insect cell transfer vector, pTMHY101, previously digested with NdeI and ScaI. Plasmid pTMHY101 contains a gene sequence that produces a peptide corresponding to the 6 EGF-like domains of thrombomodulin, amino acids 227-462, described above. pTMHY101 is described in copending U.S. application Serial No. 345,372 and in B(1)(iii) above.

Specific mutagenizing oligonucleotide primers were synthesized and used with the MUTATOR DNA Polymerase III Site-directed Mutagenesis Kit (catalog no. 200500, Stratagene, La Jolla, CA), except that, as noted elsewhere, two-strand synthesis is initiated and thrombomodulin analog genes are created in which either one or both of the methionines are converted to a nonoxidizable amino acid. Primers directing conversion to the preferred amino acids leucine, glutamine, or alanine are listed in Table 5. These primers also contain substitutions in the nucleotide sequence that provide a single cleavage site for a restriction enzyme that are diagnostic of the success of the mutagenesis, but these substitutions do not necessarily alter the corresponding amino acid sequence. In the primers shown in Table 5, the nucleotide substitutions are underlined. For example, in plasmid pTHR28, the methionine at position 388 in the native thrombomodulin protein was replaced by leucine and a single PvuII cleavage site was introduced. Other non-oxidizable amino acids that were substituted would of course also be suitable for this invention.

Purified single-stranded DNA templates were prepared using the method described by Bio-Rad (Muta-Gene Phagemid in vitro Mutagenesis, Instruction Manual, Cat. No. 170-3576, pages 33-34), although other methods known to those skilled in the art would also be suitable.

The 5' terminus of each mutagenizing primer was phosphorylated by incubating 0.5 ng/ul primer for 30 minutes at 37°C in a solution containing 2 mM rATP, 0.4 U/ul polynucleotide kinase in annealing buffer (20 mM Tris-HCl pH 7.5, 8 mM MgCl2 and 40 mM NaCl). The mixture was heat inactivated by incubating the mixture for 15 minutes at 65°C. Phosphorylation increases the mutation success rate. The phosphorylated primer was annealed to the single-stranded template by heating 100 ng of template and 2.5 ng of primer in 25 µl of annealing buffer at 65°C for 5 min, followed by cooling and annealing the mixture at room temperature for 10 min. Double-stranded DNA was prepared by extending the primer, proceeding essentially as described by Tsurushit, N., et al., (1988) Gene 62: 135-139 and O'Donnell, ME, et al., (1985) J. Biol. Chem. 260: 12875-12883. Briefly, the template/primer mixture was diluted 1:1 with 10% reassociation buffer plus 80 μg/ml bovine serum albumin, 2.5 mM dithiothreitol, 0.25 mM dNTP mix, 2 mM rATP and 1% glycerol plus 1 μg single-strand DNA binding protein. The mixture was incubated for 5 minutes at room temperature to allow the binding protein to coat the single-strand DNA template. DNA polymerase III holoenzyme (E. coli, 1.7 μl of a 50 U solution) was added and the mixture was incubated for 10 minutes at 30ºC. T4 DNA ligase (0.5 μl, 2 Weiss units) was added and the mixture was incubated for another 5 minutes at 30ºC. With this mixture E. coli was transformed and correctly mutated clones were selected based on the restriction cleavage patterns.

The same procedure can be used, for example, to generate mutants that can be expressed in mammalian cells using pTR13 (described above), which has an M13 origin of replication for the production of single-stranded DNA templates.

D. Production and purification of the recombinant protein

T25 flasks were seeded at a density of 2 x 106 with Sf9 cells in 5 ml of TMN-FH medium plus 10% FBS or Excell 400 and then infected with an isolated recombinant plaque from part B or C above. After three days, virus stocks were harvested. Flasks (30-100 ml shake flasks or 100-300 ml spinner flasks, respectively) were seeded with cells (1-1.8 x 106/ml) and infected with aliquots of virus stock corresponding to 1/50th to 1/100th of the final volume. Infected cell cultures were grown for four days and conditioned medium containing recombinant oxidation-resistant TM analog protein was harvested. The TM analogues were purified from the conditioned medium by removing cellular debris followed by five chromatography steps, namely 1) Q-Sepharose, 2) thrombin affinity, 3) gel filtration, 4) anion exchange and 5) a second gel filtration step. The gel filtration steps involve buffer re-adjustment. All chromatography steps were carried out at 4ºC.

i. Material

Some chromatography resins were purchased commercially. Q-Sepharose and Sephadex G25 were purchased from Sigma (St. Louis, MO) and Mono Q 5/5TM from Pharmacia LKB (Piscataway, NJ).

The DFP-thrombiri-agarose was prepared approximately as follows: 360 mg bovine thrombin in 100 ml of 20 mM sodium phosphate, pH 7.5, was added to approximately 100 ml of a 50% slurry of Affigel 10 resin and mixed overnight at 4°C. The Affigel 10 was prepared for use as directed by the manufacturer and equilibrated with the loading buffer. Residual activated esters were blocked by adding 100 ml of 0.1 M glycine (pH 5.6) for 1 hour at 4°C. The gel was then equilibrated with 30 mM Tris-HCl, 2 M NaCl, pH 7.5, and 20 µl of DFP was added to a final concentration of approximately 1 mM DFP. After mixing for 16 hours at 4°C, an additional 6 µl of DFP was added and mixing continued for 4 hours. The resin was then washed with 20 mM Tris-HCl, 2 M NaCl pH 7.5 and stored at 4ºC.

Thrombin activity was measured using the substrate Kabi S-2238 and it was found that > 86% of the thrombin was removed from the solution and presumably coupled to the resin, giving a final concentration of approximately 6 mg thrombin per ml resin. The enzyme activity of the DFP-treated resin was < 1% of the initial activity.

ii. Preparation of pure TM analogue peptide.

The conditioned medium was harvested and clarified by centrifugation at 1400 xg for 10 min. The pH was adjusted from approximately 6.0 to approximately 5.2 with glacial acetic acid. The adjusted medium was then applied to a column of Q-Sepharose resin. The column had previously been equilibrated with approximately four column volumes of Wash Buffer 1 (117 mM sodium acetate, 0.02% NaN3 pH 5.0). After application, the column was washed with Wash Buffer 1 and then Wash Buffer 2 (25 mM sodium acetate, 0.1 M NaCl pH 5.0) and the oxidation-resistant TM analog was then eluted with Wash Buffer 2 containing 0.3 M NaCl, pH 5.0.

The column fractions containing activity (measured in the protein C activation assay, see above) were pooled, then diluted with 0.3 M NaCl, 20 mM Tris-HCl, 0.5 mM CaCl2, 0.02% NaN3, pH 7.5. The pH of the dilution was measured and adjusted to approximately 7.5 with NaOH. The ionic strength of the pool was approximately the ionic strength of a 0.3 M NaCl solution. This adjusted pool was applied by gravity overnight to a thrombin-agarose column pre-equilibrated with the same buffer used to dilute the conditioned media. The column was washed with dilution buffer and the TM analogue was removed from the matrix with 1.5 M GuHCl, 2.0 M NaCl, 20 mM Tris HCl, 1 mM Na EDTA, 0.02% NaN3, pH 7.5.

The essentially pure TM analogue was applied to a Sephadex G25 column and recovered in 0.2% N-ethylmorpholine acetate (NEM), pH 7.0. In this step, GuHCl and NaCl are removed.

The TM analogue recovered from the Sephadex G25 column was applied to a Mono Q column (Pharmacia, 10 micron particles, quaternary amines) pre-equilibrated with 0.2% N-ethylmorpholine (NEM), pH 7.0. After washing with this buffer, the different forms were separated using a gradient from 0 to 0.4 M NaCl. Samples of each fraction were evaluated on an SDS-PAGE gel under non-reducing conditions. SDS-polyacrylamide gel electrophoresis was performed according to the method of Laemmli using 3.3% acrylamide in the stacking gel and 12.5% acrylamide in the resolving gel. Non-reduced samples were diluted with Laemmli sample solubilization buffer (50 mM Tris-HCl, pH 6.8, 25% glycerol, 2% SDS and 0.01% bromophenol blue) and applied directly to the gel. Protein standards from the LMW Calibration Kit from Pharmacia were used as MG markers and the gels were silver stained. Under these conditions, only a single band is visible when silver stained.

The fractions containing peptides with similar mobilities were pooled and then in an assay for total protein content and for activity in the Protein C activation assay as described below.

E. Assays for thrombomodulin analogues 1. Materials

Rabbit thrombomodulin, hirudin, and human protein C were from American Diagnostica. Human thrombin is available from various commercial and other sources. Bovine thrombin was obtained from Miles Labs, Dallas, Texas. D-valyl-L-leucyl-larginine-p-nitroanilide (S-2266) and D-Phe-Pip-Arg-p-nitroanilide (S-2238) were obtained from Kabi Diagnostica.

Bovine serum albumin (fraction V), human citrated plasma and APTT reagent were purchased from Sigma Chemicals. Microtiter plates were from Corning (No. 25861-96). All other reagents were of the highest possible purity.

2. Methods and results i. Protein C activation assay (color test)

In this assay, 20 µl of each of the following proteins were mixed in a microtiter plate: the thrombomodulin sample (unknown or standard), thrombin (3 nM), and protein C (1.5 µM). The assay diluent for each protein was 20 mM Tris-HCl, 0.1 M NaCl, 2.5 mM CaCl2, 5 mg/ml BSA, pH 7.4. The wells were incubated at 37ºC for 2 hours, and protein C activation was then terminated by adding 20 µl hirudin (0.16 units/µl, 370 nM) in assay diluent and incubating for a further 10 minutes.

The amount of activated protein C formed was determined by adding 100 μl of 1.0 mM S-2266 (in assay diluent) and further incubating the plate at 37ºC. The absorbance at 405 nm in each well was measured every 10 seconds for 30 minutes using a Molecular Devices microtiter plate photometer. measured. The absorbance data were stored and the absorbance change per second (curve) in each well was calculated. The absorbance change per second is proportional to the activated protein C in pmol/ml.

This ratio was determined empirically with various concentrations of fully activated protein C. Samples containing 100% activated protein C were prepared by mixing 0 to 1.5 µM protein C with 60 nM rabbit TM and 30 nM thrombin, incubating for 0 to 4 hours, adding hirudin, and measuring S2266 activity as above. The conditions under which 100% of protein C was activated were defined as plateau conditions for S2266 activity (A405/s).

One unit of activity is defined as 1 umol of activated protein C produced per ml/min under the reaction conditions defined above. However, activity values can also be expressed in comparison with native, detergent-solubilized rabbit thrombomodulin. Protein mass was derived by amino acid analysis and it was determined that 1 nmol of the TM analogue 6h/227-462 (see Table 4) has an activity equivalent to that of 1 nmol of rabbit thrombomodulin. Other TM analogues are more active than 6h/227-462 in this assay. For example, a TM analogue with the 6 EGF-like domains, with the methionine at amino acid position 388 replaced by a leucine by in vitro mutagenesis (see Table 4), has a specific activity approximately 2.2-fold higher than that of 6h/227-462.

ii. Protein C cofactor activity after exposure to oxidants

To specifically test the oxidation stability of the mutated TM analogue peptides, chloramine T (N-chloro-p-toluenesulfonamide sodium salt, Sigma) was used. The transfection culture supernatant (1 ml) containing a peptide encoded by a mutated TM gene sequence or pTMHY101 (wild type, aa 227-462) was desalted on a NAP-10 column (LKB/Pharmacia) in 1.5 ml of 0.2% N-ethylmorpholine (NEM), pH 7.0, 0.008% Tween 80, and then lyophilized and resuspended in 100 μl of the above buffer. The sample was divided in half and either 5 μl of water (control) or 5 μl of 0.1 M chloramine T (final concentration = 9.1 nM) was added to the halves. The samples were incubated for 20 minutes at room temperature and then passed through a NAP-5 column to remove residual oxidants. The desalting buffer was the diluent for the protein C assay. After contact with chloramine T, the mutated peptide retained all activity, whereas the wild-type peptide was essentially inactivated.

iii. Inhibition of activated partial thromboplastin time (APTT).

Clot formation from citrated plasma is induced by the addition of brain cephalin in ellagic acid ("APTT reagent") and calcium ions. The time required for clot formation is reproducible and increases proportionally with the addition of thrombomodulin. The APTT reagents are incubated at 37ºC before mixing, except for the citrated plasma which is kept at 4ºC.

The reaction was carried out as follows: 100 µl of Sigma citrate plasma was placed in a plastic cuvette (Sarstedt No. 67.742) and incubated for 1 minute at 37ºC; then 100 µl of Sigma APTT reagent was added and incubated for 2 minutes at 37ºC; then 100 µl of test sample (or control buffer) and 100 µl of 25 mM CaCl₂ were added and the cuvette was immediately placed in a Hewlett Packard 8451A spectrophotometer equipped with a circulating water bath so that the cuvette was kept at 37ºC during the measurement. The absorbance due to light scattering at 320 nm was measured every 0.5 seconds from the 15th to the 120th second from the start of the CaCl₂ addition. By applying the Extinction versus time results in an S-curve, where the clotting time is defined as the time at which the slope corresponding to the bend of the curve is steepest.

Ex vivo APTT assays were performed as described above, except that citrated plasma from the animal used in the in vivo experiment was used instead of commercially obtained citrated plasma.

iv. Inhibition of thrombin time (TCT) and prothrombin response (PT).

Both PT and TCT are determined using the Hewlett Packard 8452 A diode array spectrophotometer used for APTT. For the PT reaction, 20 μl of thromboplastin and 90 μl of 25 mM CaCl2 in a cuvette were added to 90 μl of either TM analogue 6h/227-462 or PBS. The mixture was incubated at 37ºC for 1 minute and then 100 μl of citrated plasma was added. After placing the cuvette in the spectrophotometer, the absorbance due to light scattering at 320 nm was measured every 0.5 seconds during the 15th to 120th second after the addition of plasma. By plotting absorbance against time, an S-curve is obtained, with clotting time defined as the time at which the slope corresponding to the bend of the curve is steepest. The TCT is evaluated in the same way. The initial reaction mixture contains 100 μl citrated plasma, 25 μl 100 mM CaCl2 and either 162.5 μl PBS or 162.5 μl of the TM analogue. After 1 minute, 12.5 μl thrombin is added. Clotting time is measured as described above.

v. Direct anticoagulant activity - Inhibition of thrombin-catalyzed conversion of fibrinogen to fibrin.

Thrombin and varying amounts of the TM analogue 6h/227-462 were incubated in microtiter plate wells for 2 minutes at 37ºC. The total initial reaction volume was 50 µl PBS + 7.5 mM CaCl₂ and 90 nM thrombin. After initial incubation 100 μl of 3.75 mg/ml human fibrinogen was added to each well and thrombin-induced fibrin formation was monitored by measuring the change in absorbance at 405 nm in a Molecular Devices Vmax spectrophotometer (Molecular Devices, Menlo Park, CA). The endpoint of the assay was the time at which 50% of the total absorbance was reached. Residual thrombin activity was determined by reference to a thrombin standard curve that linearly relates the inverse of the thrombin concentration to the clotting time. When amounts of detergent-solubilized native rabbit thrombomodulin and the TM analogue 6h/227-462, which has equivalent activity as measured by protein C Kpfactor activity, are compared in the direct anticoagulant activity assay, the TM analogue has a significantly reduced ability to inhibit thrombin-mediated conversion of fibrinogen to fibrin (approximately 1/10).

vi. Inhibition of platelet activation and aggregation.

The effect of TM analogue 6h/227-462 on thrombin activation of platelets was tested according to the methods of Esmon et al., (1983) J. Biol. Chem. 258: 12238-12242. When evaluated by this assay, TM analogue 6h/227-462 did not significantly inhibit thrombin-mediated platelet activation and aggregation.

viii. Additional measurements of the antithrombotic activity of TM.

1) Inhibition of thrombin-induced activation of factor V by the TM analogue is measured according to the procedure described by Esmon et al., J. Biol. Chem., (1982), 257: 7944-7947.

2) The inhibition of the TM analogue/thrombin complex by antithrombin III and heparin cofactor II is measured as described by Jakubowski et al. (1986).

3) Inhibition of thrombin-induced inactivation of protein S by the TM analogue is measured according to the procedure described by Thompson & Salem, J. Clin. Invest., (1986), 78(1): 13-17.

4) Inhibition of thrombin-mediated activation of factor XIII is measured according to the method of Polgar, et al., (1987) Thromb. Heamostas. 58: 140.

EXAMPLE 2. In vivo activity of a TM analogue in the rodent model of deep phlebothrombosis.

The ability of a TM analogue to abolish thrombus formation was evaluated in a modified stasis/endothelial injury-induced rat venous thrombosis model (see Maggi, A. et al., (1987) Haemostasis 17: 329-335 or Pescador, R. et al., (1989) Thrombosis Research 53: 197-201). The vena cava of an anesthetized male Spraque-Dawley rat (450 g) was surgically isolated and the animal was subsequently treated by bolus injection into the femoral artery with a thrombomodulin analogue (6h/227-462 containing the 6 EGF-like domains of native thrombomodulin), standard heparin or normal saline (0.1 ml/rat) as a control. The heparin dose was 45 units/rat. The thrombomodulin analogue dose was 100, 10, 1, 0.1 or 0.01 µg/rat. Two minutes after injection, the inferior vena cava was ligated to the left renal vein to induce stasis and the vascular endothelium was injured by gentle compression with forceps. After 10 minutes, the vena cava was harvested and examined for the presence of a thrombus, which, if present, was removed and weighed. No thrombus formation was observed in any of the animals treated with heparin or the thrombomodulin analogue (6h/227-462) at a dose of 100, 10 or 1 µg/rat, whereas the animals treated with saline and those treated with the lowest dose of the Thrombomodulin analogue (0.01 µg) had thrombi with an average weight of 14.9 mg/thrombus. In the rats treated with 0.1 µg thrombomodulin analogue, a tiny thrombus was observed that was not large enough to be removed and weighed.

The dose range used in this study was based on an in vitro APTT assay, in which 1 ug/ml thrombomodulin analogue was insufficient to prolong APTT, but the addition of 10 ug/ml resulted in a significant prolongation. Results from APTT assays performed on plasma samples from each of the treated rats showed no prolongation in the TM analogue-treated rats or in the control rats (100 ug TM analogue = 45 s, all other TM analogue doses and saline controls = 30-35 s). However, the APTT of the heparin-treated rats was significantly prolonged (100 s).

This experimental system represents a directly comparable model for deep phlebothrombosis in humans, which is characterized by vascular injury and reduced blood flow. The results described above demonstrate that very low doses of a TM analogue, which can act as a cofactor for thrombin-mediated activation of protein C, but has a significantly reduced ability to inhibit thrombin-mediated conversion of fibrinogen to fibrin, are effective in preventing thrombus formation. In addition, the absence of a prolonged APTT measured ex vivo indicates that this TM analogue does not exert a systemic effect on coagulation parameters and therefore would not increase dangerous bleeding-related side effects.

EXAMPLE 3. In vivo activity of a TM analogue in the primate model of venous and arterial thrombosis

The thrombus-inhibiting properties of thrombomodulin analogues were demonstrated in the arteriovenous shunt Baboon model evaluated using the slightly modified method of Cadroy, Y. et al., (1989) Journal of Laboratory and Clinical Medicine 113: 436-448, as described in Hanson SR and Harker, LA (1987) Thrombosis and Haemostasis 58: 801- 805. This model was chosen because of the haemostasis similarity between the baboon and man and because the arteriovenous shunt serves as a model for both arterial and venous type thrombi.

A shunt made of Silastic tubing modified with a piece of Dacron tubing (3.2 mm diameter) followed by a Teflon chamber (9.3 mm diameter) was inserted into the femoral artery of a baboon in such a way that blood from the artery flowed through the shunt and returned to the baboon via the femoral vein (see Fig. 2). The Dacron tubing provides a thrombogenic surface that stimulates the natural clotting process, particularly the deposition of platelets on the implant surface, and serves as a model for the generation of arterial thrombi, i.e., platelet-rich thrombi held together by fibrin. The chamber creates a stasis condition similar to that found in veins where blood flow velocity is reduced, and in particular mimics the region around the venous valves, thereby serving as a model for flow conditions such as those found in deep phlebothrombosis. The thrombi formed in the chambers are fibrin-rich venous-type thrombi. Venous-type thrombi also contain platelets, but less than arterial-type thrombi. The thrombus formation in the Dacron implant or chamber is evaluated by measuring platelet deposition and fibrin growth. Platelet deposition is measured by removing the platelets from the baboon, radiolabeling them with ¹¹¹indium oxine according to the method of Cadroy, Y. et al., (1989) Journal of Clinical and Laboratory Medicine 113 (4): 436-448 and then released back to the animal. A scintillation camera, such as a Picker DC 4/11 Dyna scintillation camera (Picker Corp., Nortriford, Conn.), is placed directly over the implant to directly measure the amount of radioactivity deposited by the platelets as part of a thrombus, as in Cadroy, Y. et al. As a second measure of thrombus formation, a dose of 5 uCi of baboon fibrinogen labeled with 125I is administered intravenously prior to insertion of the shunt. At the end of the experiment, the shunt is removed, washed, and stored for 30 days to allow the radioactivity of the 125indium to decay (half-life 2.8 days). Since 1111indium decays much more rapidly than 125iodine, the detectable radioactivity remaining in the shunt represents the amount of fibrin deposited as part of a thrombus. Total fibrin deposition is calculated by dividing the counts deposited per minute by the amount of clottable fibrinogen present in the baboon's blood as determined by the TCT assay. The first shunt in this series serves as a control for the second shunt.

Two shunts in series were placed in a baboon and the TM analogue (6h/227-462, see Table 4) was infused to a point between the two shunts at 7 or 8 mg/h for 1 hour. As shown in Figure 3, platelets were deposited in both the chamber and the Dacron implant in the control shunt, but platelet deposition was significantly reduced after infusion of the TM analogue into the second shunt.

From these experiments it is evident that a TM analogue which has the ability to act as a cofactor for the thrombin-mediated activation of protein C and which has a significantly reduced ability to inhibit the thrombin-mediated conversion of fibrinogen to fibrin as well as the thrombin-mediated Inhibiting platelet activation and aggregation, it can prevent both arterial-type and venous-type thrombi formation in an in vivo model. Such a TM analogue would therefore be suitable for the pharmaceutical treatment of any thrombotic disease, whether localized to the arteries or the veins.

EXAMPLE 4. Circulating in vivo half-life

The circulating half-life of several TM analogues was evaluated using the modified protocol of Bakhit, C, et al., (1988) Fibrinolysis 2: 31-36. The thrombomodulin analogue was radiolabelled with 125 iodine using the lactoperoxidase method of Spencer, S. A., et al., (1988) J. Biol. Chem. 263: 7862-7867. An amount of approximately 100,000 cpm of the labelled analogue was injected into the femoral vein of an anesthetised mouse and samples of XXX volume were taken after specified time periods. The amount of radioactivity present in each sample, corresponding to the amount of radiolabeled thrombomodulin analogue present in the circulation, was determined by counting in a gamma counter (Beckman) and the time required to reduce the amount of radioactivity in the circulation to half its original value was determined.

Three thrombomodulin analogues were evaluated by this method: 6h/227-462 (see above), 6h/227-462 pretreated with hydrofluoric acid (HF) to remove some or all of the carbohydrates, and 4t/227-462 (see Table 4 and Example 1.B.2). Treatment was carried out according to the method of Mort, AJ and Lamport, TA (1977) Analytical Biochemistry 82: 289-309. Briefly, 0.8 mg of the TM analogue (6h/227-462) was incubated in 1 ml anisole + 10 ml conc. HF for 1 hour in vacuum at 0ºC. The volatile liquid was then evaporated and the protein residue was removed by washing twice with 3 ml each 0.1 M acetic acid and then rinsed out of the reaction chamber by washing twice with 3 ml of 50% acetic acid. The combined washings were extracted with 2 ml of ethyl ether to remove anisole residues. The peptide-containing aqueous phase was desalted on a PD10 column, whereby 92% of the protein from the starting material was recovered.

As shown in the results in Table 6, treatment of the TM analogue to modify glycosylation can significantly alter its circulating half-life, either by removing the carbohydrate or by altering its composition by expression in different cell types. Table 1 Table 1 - (continued) - Table 1 - (continued) - Table 1 - (continued) - Table 2 Table 3 Table 3 - (continued) - Table 3 - (continued) - Table 4 Expression in insect cells Expression in mammalian cells Table 5 Primers for the replacement of methionine at aa 291 Primer for the replacement of methionine at AS 388 Native sequence

Table 6 Sample / Half-life in minutes

6h/227-462 2.7

HF treated 6h/227-462 7.3

4t/227-462 8.1

Claims (8)

1. Use of a purified thrombomodulin analogue having an extended in vivo circulating half-life, said analogue consisting of a polypeptide having an amino acid sequence substantially corresponding to the six epidermal growth factor-like domains, and wherein the six epidermal growth factor-like domains are modified by chemical treatment with hydrofluoric acid to remove glycosylation in the six EGF-like domains, said analogue having approximately the native ability to enhance thrombin-mediated activation of protein C compared to native thrombomodulin and having a reduced ability to inactivate thrombin-mediated conversion of fibrinogen to fibrin compared to native thrombomodulin, for the manufacture of a pharmaceutical composition. 2. Use of a thrombomodulin according to claim 1, wherein the thrombomodulin analogue is soluble in aqueous solution. 3. Use of a thrombomodulin according to claim 2, wherein the thrombomodulin analogue is modified at the sugar residues of the six epidermal growth factor-like domains from position 227-462. 4. Use of a thrombomodulin according to claim 2, wherein at least one amino acid residue forming a glycosylation site in the six epidermal growth factor-like domains has been deleted. 5. Use of a thrombomodulin according to claim 1, wherein the analog has 80% or less of the ability of native thrombomodulin to inactivate thrombin-mediated conversion of fibrinogen to fibrin, relative to native detergent-solubilized rabbit thrombomodulin, on equal units of protein C activation. 6. Use of a thrombomodulin according to claim 1, wherein the analogue is compared to native, Detergent-solubilized rabbit thrombomodulin has, with respect to equal units of protein C activation, 50% or less of the ability of native thrombomodulin to inactivate the thrombin-mediated conversion of fibrinogen to fibrin. 7. Use of a thrombomodulin according to claim 1, wherein the analogue is oxidation resistant. 8. Method for prolonging the in vivo circulating half-life of a thrombomodulin analogue by removing glycosylation in the six EGF-like domains.